Download Application of Artificial Intelligence in Finance

Application of Artificial Intelligence in Finance
EP-497 First stage B.Tech Project
Submitted in Partial fulfillment of the requirements for the B.Tech Degree
Submitted
by
Avnish Nainawatee
Roll No 02026013
Under the guidance of
Prof. R. M. Sonar
(SJM, School of Management)
and
Prof N.Nambudripad
(Department of Physics, IIT Bombay)
Department of Physics
Indian Institute of Technology, Bombay
November 2005
Acceptance Certificate
Department of Physics
Indian Institute of Technology, Bombay
This is to certify that this report entitled “Application of Artificial Intelligence in
Finance” is based on the work carried out by Avnish Nainawatee (02026013)
under my guidance.
th
Date: 16 November 15, 2005
Prof R. M. Sonar
ACKNOWLEDGEMENT
th
Date: 16 November, 2005
I would like to take this opportunity to express my deep sense of gratitude to my guide Prof.
R.M Sonar who not only gave me proper guidance on the topic throughout the semester but also
kept encouraging me from time to time. I would also like to thank my Co-Guide Prof. N.
Nambudripad for his constant support and Godse Manish Dadasaheb for his suggestions.
Avnish Nainawatee
02026013
November 2005
Contents
Acceptance Certificate……………………………………………………………………..ii
Acknowledgements……………………………………..………….…………...…………iii
Contents……………………………………………………...…..………………...………iv List
of figures ……………………………………….…….…….………………..……….vi
Abstract…………………………………………………………………………...……....vii
1. Introduction……………………………………………………..………………...…...1
1.1 Artificial Intelligence...……………………………………………………..…1
1.2 Expert System (E.S.)…………………………………………… …..……..….2
1.3 Genetic Algorithm (GA)…………….,………….… …………………………..2
1.4 Knowledge Base ……….……………………….……………………………..2
1.5 Use of AI for Investment Portfolio ………………………………………..…...3
2. Expert System…………………………….…………..…………………………………4
2.1 Characteristic of Expert System …………….……………………………………4
2.2 Building Blocks for Expert System ………………..……………….………….....4
2.2.1 Knowledge base ………………………..………………….………….5
2.2.2 Reasoning engine: …………………………………………………….5
2.2.3 Knowledge acquisition subsystem ……………………………………5
2.2.4 Explanation subsystem …………………………………….………….5
2.2.5 User interface …………………………………………………………6
2.3 Expert System Algorithm
2.3.1 Forward chaining …………………………………………………….6
2.3.2 Backward chaining…………………………………..……………….6
2.4 Reasoning with Uncertainty…………………………………….……………….8
2.5 Use of Expert System………………………………………………..…………..8
3. Case-Based Reasoning……………………………………………………..……………9
3.1 CBR Approach…………………………………………………………………….9
3.1.1 Textual CBR approach ………………………………………………..9
3.1.2 Conversational CBR approach……………………………….………10
3.1.3 Structural CBR approach…………………………………………….10
3.2 Working of CBR: 4 –R approach………………………………..……….………10
3.2.1 Retrieving………………………………….…………………….…...11
3.2.2 Reusing………………………………….…………………….……..12
3.2.3 Revising………………………………….…………………….…….12
3.2.4 Retaining………………………………….…………………….……12
3.3 Similarity Assessment………………….…………………………………. …….12
4. Genetic Algorithm……………………………………………………………….…....13
4.1 Operation of a Gentic Algorithm ………………………………………….........13
4.2 Observations about Genetic Algorithm …………………………………………15
5. Investment Alternative…………………………………………...……………………17
5.1 Non-Marketable financial Assets…………………………………………...…....17
5.1.1 Bank Deposits.…………………………………………...……...…...17
5.1.2 Post Office Time Deposit (POTDs) ………………………………... 18
5.1.3 Monthly Income Scheme of Post Office (MISPO) ………………… 18
5.1.4 Kisan Vikas Patra (KVP) …………………………………………... 18
5.1.5 National Saving Certificate…………………………….…………… 18
5.1.6 National Saving Scheme(NSS) ………………………….…………. 19
5.1.7 Company Deposits…………………………………….……...…….. 19
5.1.8 Employees Provident Fund Scheme………………………….…….. 20
5.1.9 Public Provident Fund Scheme 20
5.2 Money Market Instruments………………………………………………….….. 20
5.2.1 Treasury Bills…………………………………………...……….….. 20
5.2.2 Certificates of Deposits…………………………………………....... 21
5.2.3 Commercial Paper…………………………………………...……… 21
5.2.4 Repos…………………………………………...…………………… 21
5.3 Bonds…………………………………………...………………………………. 21
5.3.1 Government Securities…………………………………………....... 22
5.3.2 RBI Relief Bonds…………………………………………............... 22
5.3.3 Infrastructure Bonds…………………………………………...…… 22
5.3.4 Private Sector Debenture………………………………………….... 23
5.3.5 Public Sector Undertaking Bonds. …………………………………. 23
5.3.6 Preference Shares 23
5.4 Financial Derivatives: Futures…………………………………………...…….. 24
5.5 Financial Derivatives: Options…………………………………………...……. .24
5.6 Life Insurance…………………………………………...……………………… 24
5.7 Mutual Funds…………………………………………...……………………… .26
5.7.1 Mutual Funds Open end /Close end scheme……………………….. 26
5.7.2 Mutual Funds: Equity Schemes and Debt Scheme…………..…….. .27
6 Intelligent Systems for Investment Portfolio………………………………………… 28
6.1 Combining Expert Systems and CBR…………………………………………... 29
6.2 Combining CBR and Genetic Algorithm……………………………………….. 29
7 Conclusions…………………………………………...………………………………... 30
7.1 Possible route for the Second stage…………………………………………...... 30
Reference…………………………………………...……………………………………. 31
List of Figures
2.1: Structure of Expert System……………………………………………………………5
2.2: Working of Backward Chaining………………………………………………...……7
3.1: Working of CBR: 4 –R approach ……………………………………………..…….11
6.1: Model for Applying ES, CBR and Genetic Algorithm for investment Portfolio...…28
Abstract
In the past few years, Artificial Intelligence systems have found new applications in the
finance industry, an area dominated by traditional analytic techniques. AI algorithms are used for
optimization and finding solutions to problems in finance.
In this report first we discuss overview of Expert System, Case Base Reasoning and Genetic
Algorithm which we will be using to determine Investment Portfolio. Then, we have discussed
about various Investment Options and there important feature. Finally we propose structure of
Hybrid Intelligent system which combines all three approaches mentioned above for determining
Investment Portfolio.
Chapter 1
Introduction
In this Chapter we will discuss need of Artificial Intelligence in finance. We will also introduce
various AI techniques used in this project and how we have integrated them for application in
finance
1.1 Artificial Intelligence (AI) and Finance
Artificial intelligent represent use of advance computer intelligence instead of traditional
analytic techniques. This is as efficient method of handling large knowledge base and arriving at
intelligent solution to problem using that knowledge [1]
AI has wide range and significant application in finance [2,3]. It can help consumers source
the best financial options, allowing them to manage their money more efficiently. Equipped with
details of present financial products such as interest rates of savings accounts or mortgages or
loans, agents will be able to identify the products that best meet a user's requirements. It helps in
exploiting data so that smarter business decisions can be made in less time and/or at lower cost.
In this project we have used Expert system Case Base Reasoning (CBR) and Genetic
algorithm to determine the financial portfolio for Investment.
1.2 Case Base Reasoning (CBR)
A case-based reasoning solves new problems by using or adapting solutions that were used to
solve old problems [4]. It offers a reasoning paradigm that is similar to the way many people
routinely solve problems.
The current problem is matched against the cases in the case base, and similar cases are
retrieved. The retrieved cases are used to suggest a solution which is reused and tested for
success. If necessary, the solution is then revised. Finally the current problem and the final
solution are retained as part of a new case.
1.3 Expert Systems (ES)
An expert systems are programs made up of a set of rules that analyze information supplied
by the user of the system about a specific class of problems, as well as provide analysis of the
problem [5]. Expert system provides expert quality advice, diagnoses and recommendations for
given real world problems.
Expert systems are constructed by obtaining this knowledge from a human expert and coding
it into a form that a computer may apply to similar problems.This reliance on the knowledge of a
human domain expert for the system's problem solving strategies is a major feature of expert
systems.
1.4
Genetic Algorithm (GA)
A genetic algorithm (GA) is a search technique used in to find approximate solutions to
optimization and search problems. Genetic algorithms are a particular class of evolutionary
algorithms that use techniques inspired by evolutionary biology such as inheritance, mutation,
natural selection, and recombination [6,7]
This technique is used to determine most optimal solution to the problem.We have used this
algorithm to determine the amount to be invested in various options given by CBR retrival which
maximises the profit
1.5
Knowledge Base
Knowledge Base is store of factual and heuristic knowledge. An AI tool provides one or
more knowledge representation schemes for expressing knowledge about the application domain.
We will be using the special model of AI algorithm for this project which uses both expert
system and Case Base reasoning to obtain the final portfolio. Rules that are used for Expert
System are also used by the case-based component to do case indexing and case adaptation.
We have explored the various investment options and developed the sample Cases which
represent the knowledge base. Important case parameters are Risk, rate of interest, time of
investment, liquidity, loan against investment, tax exemption.
1.6 Use of Intelligent System for Investment Portfolio
Expert system Case Base Reasoning (CBR) and Genetic algorithm are combined to form
Intelligent System in following project to determine the portfolio according to input specification
of the user.
First Expert system is use to take then input from the user and produce some relevant input
case structures. These are input to CBR which gives suitable investment option which are also
stored in form of cases.
Then we use Genetic Algorithm to determine the division of allotted fund in the given
portfolio which will results in maximum return or balance the risk profile based according to
input specification.
1.7
Organization of Report
Report is organized in to 6 Chapters. Chapter 2 deals with Expert System and its component.
Chapter 3 expressed about Case base reasoning and its working. Chapter 4 discuss about Genetic
Algorithm and how it is used for optimization. Chapter 5 gives knowledge base of investment
options. Chapter 6 addresses how project will be using algorithm and knowledge base. Chapter 7
gives the conclusion of the study.
Chapter 2
Expert Systems
Expert system provides expert quality advice, diagnoses and recommendations given real
world problems. Expert systems are constructed by obtaining this knowledge from a human
expert and coding it into a form that a computer may apply to similar problems
2.1 Characteristic of Expert System
Computer-based expert systems match the performance of a human expert more closely than
checklists, flow charts and decision tables [8]. While these devices are certainly goal-oriented
they may not compare favorably with live interaction or expert systems when efficiency,
adaptively, use of imperfect information and explanation of reasoning are important.
Expert system should be:
1. Goal oriented.
2. Efficient
3. Adaptive
4. Able to work with imperfect information
5. Should justify its recommendations by explaining their reasoning
2.2 Building Blocks for Expert System
Expert System consists of knowledge base, Inference Engine and User Interface. Each
component play important role in functioning of Expert System. Stored knowledge is used with
rules to derive inference about the problem [9, 10].
2.2.1 Knowledge base
Knowledge base is a store of factual and heuristic knowledge. An ES tool provides one or
more knowledge representation schemes for expressing knowledge about the application domain.
Some tools use both frames (objects) and IF-THEN rules. In PROLOG the knowledge is
represented as logical statements.
2.2.2 Reasoning engine
Reasoning engine is Inference mechanisms for manipulating the symbolic information and
knowledge in the knowledge base to form a line of reasoning in solving a problem. The inference
mechanism can range from simple backward chaining of IF-THEN rules to case-based
reasoning.
Figure 2.1: Structure of Expert System [11]
2.2.3 Knowledge acquisition subsystem
Knowledge acquisition subsystem is subsystem to help experts build knowledge bases.
Collecting knowledge needed to solve problems and build the knowledge base continues to be
the biggest bottleneck in building expert systems.
2.2.4 Explanation subsystem
Explanation subsystem is a subsystem that explains the system's actions. The explanation can
range from how the final or intermediate solutions were arrived at to justifying the need for
additional data.
2.2.5 User interface
User interface is a means of communication with the user. The user interface is generally not a
part of the ES technology, and was not given much attention in the past. However, it is now
widely accepted that the user interface can make a critical difference in the perceived utility of a
system regardless of the systems.
2.3 Expert System Algorithm
Expert system uses two algorithms to derive inference about particular problem these are
forward chaining and backward chaining [13].
2.3.1 Forward chaining
Forward Chaining method begins with set of known facts or attributes and applies these values
to rules that use them in their premise. Any rules that are proven true fire and produce additional
facts that are again applied to relevant rules. The process continues until no new facts are
produced or a value for the goal is obtained. This approach works well when it is natural to
gather multiple facts before trying to draw any conclusions and when there are many possible
conclusions to be drawn from the facts.
2.3.2 Backward chaining
An alternative approach begins with a rule that could conclude the goal for the consultation,
tries to obtain values for the attributes used in the rule's premise, then backtracks through
additional rules if necessary to determine a value of the goal attribute. When there are many
attributes employed in many rules, the backward chaining mechanism produces a more efficient
interview than forward chaining because it will not be necessary to ask the user to input values of
all of the facts. We will be using Backward Chaining Algorithm in project.
Backward chaining systems are described as hypothesis driven because they operate by
selecting successive rules that can determine the value of a goal or sub-goal. This value becomes
the hypothesis to be proven or disproved. Figure 2 show the working of Backward Chaining
algorithm .Once an attribute is identified inference engine puts aside the rule it is working with
and sets up a new goal and sub-goal to prove the If part of this rule. Then knowledge base is
searched again for the rules that can prove the sub-goal. The inference engine repeats the process
of stacking the rule until no rules are found in the knowledge base to prove the current goal.
Figure 2: Working of Backward Chaining [13]
2.4 Reasoning with Uncertainty
Knowledge is almost always incomplete and uncertain. To deal with uncertain knowledge, a
rule may have associated with it a confidence factor or a weight. The set of methods for using
uncertain knowledge in combination with uncertain data in the reasoning process is called
reasoning with uncertainty. An important subclass of methods for reasoning with uncertainty is
called "fuzzy logic" [9].
2.5 Features and Observation: Expert System
Expert System applications has many benefit to the end users, here we discuss about few
advantages of ES.
The sequence of steps taken to reach a conclusion is dynamically synthesized with
each new case. It is not explicitly programmed when the system is built.
Expert systems can process multiple values for any problem parameter. This permits
more than one line of reasoning to be pursued and the results of incomplete (not fully
determined) reasoning to be presented.
Problem solving is accomplished by applying specific knowledge rather than specific
technique. This is a key idea in expert systems technology. It reflects the belief that
human experts do not process their knowledge differently from others, but they do
possess different knowledge. With this philosophy, when one finds that their expert
system does not produce the desired results, work begins to expand the knowledge
base, not to re-program the procedures.
Advantage of expert systems over traditional methods of programming is that they
allow the use of confidences or reasoning with uncertainty.
Preservation of scarce expertise. ESs are used to preserve scarce know-how in
organizations, to capture the expertise of individuals who are retiring, and to preserve
corporate know-how so that it can be widely distributed to other factories, offices or
plants of the company.
Improved quality of decision making. In some cases, the quality or correctness of
decisions evaluated after the fact show a ten-fold improvement.
Speed-up of human professional or semi-professional work -- typically by a factor of
ten and sometimes by a factor of a hundred.
Chapter 3
Case-Based Reasoning
Case-based decision support is both a methodology that models human reasoning and
thinking, and a methodology for building intelligent computer systems. Cases are stored in
memory, and the case-based decision support system analyzes them to retrieve similar cases
from memory for decision making. This is the principle underlying case-based reasoning
technologies [4]. Case-based decision support can also analyze cases to extract patterns and
discover knowledge hidden in data. Case-based information system helps to exploit data so that
smarter business decisions can be made in less time and/or at lower cost.
3.1 CBR Approach
Case-based reasoning means learning from previous experiences. Given the fact that this is a
very general approach to human problem-solving behavior, it is more than natural that there are
different approaches for implementing this process on computer systems [4]. In commercial
CBR systems, there are three main approaches that differ in the sources, materials, and
knowledge they are Textual CBR approach, Conversational CBR approach and Structural CBR
approach
3.1.1 Textual CBR approach
In Textual CBR cases are recorded as free text. The CBR retrieval engine then uses various
keyword matching techniques to retrieve cases. It is very useful in domains where large
collections of know-how documents already exist and the intended user is able to immediately
make use of the knowledge contained in the respective documents.
The approach is well suited when there are couples of hundred cases and when each case has
a short description Textual CBR retrieves a large number of cases are irrelevant. The cost for
controlling the quality of textual CBR is high.
3.1.2 Conversational CBR approach
The principle is to capture the knowledge contained in customer/agent conversations. A case
is represented through a list of questions that varies from one case to the other. There is no
domain model and no standardized structure for all the cases.
The conversational CBR approach is very useful for domains where a high volume of
simple problems must be solved again and again. The system guides the agent and the customer
with predefined dialogs. The conversational approach is well suited for applications in which
only a few questions are needed for decision making.
3.1.3 Structural CBR approach
The structural CBR approach relies on cases that are described with attributes and values
that are pre-defined. In different structural CBR systems, attributes may be organized as flat
tables, or as sets of tables with relations, or they may be structured in object-oriented manner.
The structural CBR approach is useful in domains where additional knowledge, beside cases,
must be used in order to produce good results. The domain model insures that new cases are of
high quality and the maintenance effort is low. This approach always gives better results than the
two others, but it requires an initial investment to produce the domain model. We will be using
the structural CBR approach for our project.
The idea underlying the structural CBR approach is to represent cases according to a
common structure called the domain model. The domain model specifies a set of attributes (also
called features) that are used to represent a case. This set of attributes is structured by the domain
model. In a relational model, there are several tables that are indexed by a primary key. Each
case has a unique key, and the case description is distributed in several tables. The distribution is
defined by the relations of the database. In an object-oriented database, objects are decomposed
into sub-objects.
3.2 Working of CBR: 4 –R approach
To solve a current problem: the problem is matched against the cases in the case base, and
similar cases are retrieved. The retrieved cases are used to suggest a solution which is reused and
tested for success. If necessary, the solution is then revised. Finally the current problem and the
final solution are retained as part of a new case [15, 16].
Figure 3.1: Working of CBR: 4 –R approach [16]
3.2.1 Retrieving
Retrieving a case starts with a (possibly partial) problem description and ends when a best
matching case has been found. The subtasks involve:
1. Identifying a set of relevant problem descriptors;
2. Matching the case and returning a set of sufficiently similar cases (given a similarity
threshold of some kind); and
3. Selecting the best case from the set of cases returned
3.2.2 Reusing
Reusing the retrieved case solution in the context of the new case focuses on: identifying the
differences between the retrieved and the current case; and identifying the part of a retrieved case
which can be transferred to the new case. Generally the solution of the retrieved case is
transferred to the new case directly as its solution case.
3.2.3 Revising
Revising the case solution generated by the reuse process is necessary when the solution
proves incorrect. This provides an opportunity to learn from failure.
3.2.4 Retaining
Retaining the case is the process of incorporating whatever is useful from the new case into
the case library. This involves deciding what information to retain and in what form to retain it;
how to index the case for future retrieval; and integrating the new case into the case library.
3.3 Similarity Assessment
Similarity Assessment refers to finding the similarities between the input query and present
cases hence most appropriate case could be retrieved. The retrieval of cases that are most likely
to be useful in the solution of a target problem relies on an accurate assessment of their similarity
to the target problem [17,21]. Improving retrieval performance through the development of more
effective approaches to similarity assessment has been the focus of a considerable amount of
research in CBR.
In some applications of CBR, it may be adequate to assess the similarity of the stored cases
in terms of their surface features. The surface features of a case are those that are provided as
part of its description and are typically represented using attribute-value pairs. In other
applications, it may be necessary to use deep or derived features obtained from the case
description by inference based on domain knowledge. In yet other applications, cases are
represented by complex structures (such as graphs or first-order terms) and retrieval requires an
assessment of their structural similarity.
In approaches to retrieval based on surface features, the similarity of each case to the target
problem, typically represented as a real number in the range from 0 to 1, is computed according
to a given similarity measure. Usually the retrieved cases are the k cases that are most similar to
the target problem, an approach often referred to as “k nearest neighbour” retrieval or simply kNN.(retrival_overview)
3.4 Features and Observations: Case Base Reasoning
Case Base Reasoning has several advantages over Expert System or rule based reasoning. In this
section we discuss some of salient features of Case Base Reasoning.
Case-based reasoning differs from traditional rule-based systems in the sense that
knowledge is not represented in rules but in examples. Case-based reasoning builds on
the idea that human expertise is not composed of formal structures like rules, but of
experience: human expert reasons by relating a new problem to previous ones
Case-based reasoning now amounts to reasoning by comparing a new problem with a set
of stored previous problems with their solution. The solution to the new problem is
constructed by retrieving similar problems from memory and adapting their associated
solutions to apply to the new problem.
Case-based reasoning has several advantages over reasoning with rules. The main
advantage is that it is relatively easy to set up a knowledge base. While experience has
shown that it generally is very difficult to capture knowledge on a problem domain in a
set of rules, examples of problems in this domain with their associated solution are often
readily available or can easily be acquired.
Another advantage is that case-based reasoning can be used in problem domains that are
not well understood.
Expanding a case-based reasoning system amounts to adding new appropriate examples
to the set of cases. Expanding a rule-based system on the other hand is much more
difficult: adding one rule often means rewriting a large part of the rules.
A major problem in case-based reasoning however, resides in the retrieval of cases that
are sufficiently similar to a new problem at hand. The difficulty with this approach is that
it is hard to find a similarity measure that actually gives high values to cases that are
similar to the new problem. Several different similarity measures have been designed,
mostly with a specific domain of application in mind.
Chapter 4
Genetic Algorithm
A genetic algorithm (GA) is a search technique used to find approximate solutions to
combinatorial optimization problems. Genetic algorithms are a particular class of evolutionary
algorithms that use techniques inspired by evolutionary biology such as inheritance, mutation,
natural selection, and recombination (or crossover).
In Genetic algorithms is implemented as a computer simulation in which a population of
abstract representations (called chromosomes) of candidate solutions (called individuals) to an
optimization problem evolves toward better solutions. Traditionally, solutions are represented in
binary as strings of 0s and 1s, but different encodings are also possible [6,7]. The evolution starts
from a population of completely random individuals and happens in generations. In each
generation, the fitness of the whole population is evaluated, multiple individuals are
stochastically selected from the current population (based on their fitness), modified (mutated or
recombined) to form a new population, which becomes current in the next iteration of the
algorithm
After the application of the CBR we will get some concern cases satisfying the input
condition .We now have to the rationing of money in various alternative. This rationing is done
by Use of genetic algorithm
4.1 Operation of a Gentic Algorithm
An individual, or solution to the problem to be solved, is represented by a list of parameters,
called chromosome or genome. Chromosomes are typically represented as simple strings of data
and instructions, although a wide variety of other data structures for storing chromosomes may
also be used[24,25].
Initially several such individuals are randomly generated to form the first initial population.
The user of the algorithm may seed the gene pool with "hints" to form an initial population of
possible solutions
During each successive generation, each individual is evaluated, and a value of fitness is
returned by a fitness function. The pool is sorted, with those having better fitness (representing
better solutions to the problem) ranked at the top. Notice that "better" in this context is relative,
as initial solutions are all likely to be rather poor.
The next step is to generate a second generation population of organisms, based on the
processes of selection and reproduction of selected individuals through genetic operators;
crossover (or recombination), and mutation. For each individual to be produced, a pair of parent
organisms is selected for breeding. Selection is biased towards elements of the initial generation
which have better fitness, though it is usually not so biased that poorer elements have no chance
to participate, in order to prevent the population from converging too early to a sub-optimal or
local solution. There are several well-defined organism selection methods; roulette wheel
selection and tournament selection are popular methods.
Following selection, the crossover (or recombination) operation is performed upon the
selected chromosomes. Commonly, genetic algorithms have a probability of crossover, typically
between 0.6 and 1.0, which encodes the probability that two selected organisms will actually
breed. Organisms are recombined by this probability. Crossover results in two new child
chromosomes, which are added to the next generation population. The chromosomes of the
parents are mixed during crossover, typically by simply swapping a portion of the underlying
data structure (although other, more complex merging mechanisms have proved useful for
certain types of problems). This process is repeated with different parent organisms until there
are an appropriate number of candidate solutions in the next generation population
The next step is to mutate the newly created offspring. Typical genetic algorithms have a
fixed, very small probability of mutation on the order of 0.01 or less. Based on this probability,
the new child organism's chromosome is randomly mutated, typically by flipping bits in the
chromosome data structure.
These processes ultimately result in the next generation population of chromosomes that is
different from the initial generation. Generally the average fitness will have increased by this
procedure for the population, since only the best organisms from the first generation are selected
for breeding. The entire process is repeated for this second generation: each organism is
evaluated, the fitness value for each organism is obtained, pairs are selected for breeding, a third
generation population is generated, etc.
This generational process is repeated until a termination condition has been reached.
Common terminating conditions are
• Fixed
number of generations reached
• Allocated
• An
budget (computation time/money) reached
individual is found that satisfies minimum criteria
•
The highest ranking individual's fitness is reaching or has reached a plateau such that
successive iterations do no longer produce better results
• Manual
inspection
• Combinations
of the above
4.2 Features and Observations:Genetic Algorithm
There are several general observations about the generation of solutions via a genetic algorithm
•
Unless the fitness function is handled properly, GAs may have a tendency to converge
towards local optima rather than the global optimum of the problem.
•
Operating on dynamic data sets is difficult, as genomes begin to converge early on
towards solutions which may no longer be valid for later data. Several methods have been
proposed to remedy this by increasing genetic diversity somehow and preventing early
convergence, either by increasing the probability of mutation when the solution quality
drops (called triggered hypermutation), or by occasionally introducing entirely new,
randomly generated elements into the gene pool (called random immigrants).
•
Selection is clearly an important genetic operator, but opinion is divided over the
importance of crossover versus mutation. Some argue that crossover is the most
important, while mutation is only necessary to ensure that potential solutions are not lost.
Others argue that crossover in a largely uniform population only serves to propagate
innovations originally found by mutation, and in a non-uniform population crossover is
nearly always equivalent to a very large mutation (which is likely to be catastrophic).
• GAs
•
can rapidly locate good solutions, even for difficult search spaces.
For specific optimization problems and problem instantiations, simpler optimization
algorithms may find better solutions than genetic algorithms (given the same amount of
computation time). GA practitioners may wish to try other algorithms in addition to GAs.
•
GAs cannot effectively solve problems in which there is no way to judge the fitness of
an answer other than right/wrong, as there is no way to converge on the solution. These
problems are often called "needle in a haystack" problems.
•
As with all current machine learning problems it is worth tuning the parameters such as
mutation probability, recombination probability and population size to find reasonable
settings for the problem class you are working on. A very small mutation rate may lead to
genetic drift (which is non-ergodic in nature) or premature convergence of the genetic
algorithm in a local optimum. A mutation rate that is too high may lead to loss of good
solutions. There are theoretical but not yet practical upper and lower bounds for these
parameters that can help guide selection.
•
How the fitness function is evaluated is an important factor in speed and efficiency of
the algorithm
Chapter 5
Investment Alternative
Investments are the driving force of any market of the world. They are the sours of funds
which drives every production activity. Investment could have many purposes prominent one are
earning short term or long term profit, social security, Tax exemption etc. In this section
mentions about the various options available for investment. It also specifies features of various
investments. These features such as risk, rate of return, time of investment, Tax exemption, Loan
against the investment will be used as attributes in case formation. Investment option can be
broadly classify as [26]
Non-Marketable Financial Assets
Equity Shares
Bonds
Mutual Fund Scheme
Real State
Money Market Instrument
Life Insurance Policies
5.1 Non-Marketable financial Assets
Large Portion of financial assets of individual investor is held in form of non-marketable
assets. Distinguish feature of is that they represent personal transaction
5.1.1 Bank Deposits
This is simples investment avenue There are various kind of bank account : Current account,
Saving account, Fixed deposit account. Its features:
Very Safe :Regulations of Reserve Bank of India
Ceiling on Interest Payable
Interest Rate Depends on term of Deposit
Interest paid Quarterly
Loan can be raised against the bank deposits
Tax-exempt section 80L
5.1.2 Post Office Time Deposit (POTDs)
Similar to fixed deposit of commercial banks, POTD’s have following features
Similar to Deposit of the commercial bank
Deposit can be made in multiple of Rs 50
Higher interest rate 7.5% and 9% , compounded quarterly
Half yearly payment of interest
No withdrawal up to 6 month
Can be pledged for loan
Interest income upto Rs 9,000 exempt under section 80L of the Income Tax Act, 1961
No tax is deducted at source*
5.1.3 Monthly Income Scheme of Post Office (MISPO)
IT is a popular scheme of Post office meant to provide monthly income to the investor. It
feature
Provide monthly income to the investor
Term of 6 years with Interest Rate of 8%
Interest income is tax exempt (Section 80L)
No tax deduction at source
Facility of premature withdrawal
Maximum Rs 3,00,000 for single account and 6,00,000 for joint account
5.1.4 Kisan Vikas Patra (KVP)
A scheme of the post office, the Kisan Vikar Patra has the following features:
• Minimum amount of investment Rs 1000
• The investment doubles in 8 years and 7 months
• No tax deduction at sourse
• KVP can be pledge as security for loan
• Withdrawal facility after 21/2 years
5.1.5 National Saving Certificate
Issued by the post office, The National Saving Certificate offers the following features:
In denomination of Rs 100, Rs 500, Rs 1000, Rs 5000, Rs 10000
Term of 6 year. Compounded rate 8%
NSC investment is eligible for tax exemption .. Under certain limit
Interest tax :Tax exemption under section 80C
No tax deduction at sourse
Can be pledged as collateral for loans
5.1.6 National Saving Scheme(NSS)
Essential a scheme for deferring tax payment, the National Saving Scheme(NSS) has
following feature:
Scheme for deferring the Tax Payment
Account opened at post offices.
Interest of 9% per annum compounded
Amount up to Rs 40,000 tax rebate under Section 88
Balance exempt from wealth tax
One withdrawal of stipulated amount after 3 year permitted
Amount withdraw is taxable
After death, successor not liable to be taxed
No defined maturity period can be closed after 3 year
5.1.7 Company Deposits
Many companies, large and small, solicit fixed deposits from the public. Fixed deposits
mobilized manufacturing companies are regulated by the Company Law Board and fixed
deposits mobilized by the finance companies are regulated by Reserve Bank of India. Key
feature of company deposits are as follows:
Maturity :Manufacturing Company – 3 years
:Finance Company – 25 month to 5 years
Interest Rate : Can be annually, semi-annually, quarterly, monthly, accumulated
payment
Security : Fixed deposits represent unsecured loan
Credit Rating: Rating given by CRISIL which classify according to safety of deposit.
FAAA (Highest Safety) ,FAA(High Safety), FA(AdequateSafety),
FB(Inadequate Safety), FC(High Risk) and FD(Default)
Tax Treatment : Interest if fully taxable
Front End Benefit : Company compensate Broker thus passed on to depositors
Other Incentive : Premature Withdrawal ,Free Personal Accident, loan Facilities
,Allotment of equity shares and debenture
5.1.8 Employees Provident Fund Scheme
A Major vehicle of savings for salaried employees, the employee provident fund scheme has
following features:
Important saving option for Salaried Employee
Both employer and employee contribute monthly
Employer contribution Taxable, employee rebate under section 88
Interest exempted from the Tax but accumulated
Exempt from the wealth Tax
Can take Loan again the Provident Fund
Interest rate of 9.5%
5.1.9 Public Provident Fund Scheme
One of the most attractive investment avenues available in India, the Public Provident Fund
(PPF) scheme has following features:
PPF account in State Bank of India.
Period of 15 year with contribution of 16 years
Minimum of Rs 100 and Maximum of Rs 70,000
Deposit qualify for U/s 80C and 10 of the IT Act
Interest earn is 8% and is exempt from tax
Exempt from the wealth Tax
Can take loan from third year to sixth year after PPF account. Loan 25% of preceding
financial year.
After sixth year one withdrawal per year.Less than 50% of preceding year balance or end
of 4th preceding year*
Account can be continued after 15 year with block period of 5 year*
5.2 Money Market Instruments
Debt instruments with maturities less than one year at the time of issue are called money
market instruments. These instruments are highly liquid and have negligible risk. Major money
market instruments are Treasury Bills, Commercial Paper, Certificates of Deposits, and Repos
5.2.1 Treasury Bills
Most important Money Market Instrument. It has following features:
Obligation of Government of India tenor of 91 days and 364 days
Do not carry any Explicit Interest Rate sold at Discount and redeemed at Par.
Yield function of Discount and the Period of Maturity
Low Yield
Very active Secondary Market and Transacted readily
5.2.2 Certificates of Deposits
Certificates of Deposits are short term deposits can be transferred to one party to another. It
has following important features:
Issued by financial institution and banks
Principle investors in CDs Banks, Financial Instruments, Corporate, and mutual funds
Issued at a discount with Maturity of 3 months to 1 year
Available in Various Denomination
CDs are generally Risk-Free
Higher Rate of Interest compare to Treasury Bills or Term Deposits
5.2.3 Commercial Paper
A commercial paper represents short-term unsecured promissory notes issued by firm that are
generally considered to be financially strong. A commercial paper usually has maturity period of
90 to 180 days. It is sold on Discount and Redeemed at Par hence yield function of Discount and
Period of Maturity.
5.2.4 Repos
Repos is abbreviation for Repurchase Agreement or Ready Forward. Repos involve
simultaneous “Sale and Repurchase “agreement. If a Party in needs of short term funds sells its
securities at certain price and purchase back at slightly higher price. Difference in Sales and
purchase price represent interest cost of that time. It is very convenient instrument for short term
investment. It is Safe and earns predetermined return.
5.3 Bonds
Bonds or debentures represent long-term debt instruments. The issuer of a bond promises to
pay a stipulated stream of cash flows. This comprises of periodic interest payments over a period
of time. This section discuss various fixed income instruments
5.3.1 Government Securities
These are Debt Securities issued by Central Government, State Government and other
Government Agencies. They have following features:
Carry the name of holder and register at Public Debt office(PDO)
PDO pays interest on specified date of payment
Maturity Range from 3 – 20 years
Low Interest Rate and Long Maturity Date
5.3.2 RBI Relief Bonds
A popular instrument for earning tax-exempt income, RBI bonds have the following features:
Minimum Rs 1000 .Maturity Period of 5 years
Cumulative option and Non Cumulative Option
Cumulative option, issued at a face value of Rs 1,000 redeemed at Rs 1,516.
Non Cumulative :Interest 8.5% compounded half yearly
Interest earned Exempt from the Tax, Bond exempt from the Wealth Tax
No benefit of tax on money invested
Bonds issued in from of Bond Ledger Account and Promissory Notes
Bonds are transferable
Bonds can be offered as Security for the Loans
5.3.3 Infrastructure Bonds
This is money raised by government for infrastructural development. They have following
feature:
Provide tax-saving benefits under Section 88.
Face value of Rs 5,000 for 3 years @ 9.00% interest payable annually.
Deep Discount Bonds: Face value of Rs 6,600, available for Rs 5,000,issued for 3
Infrastructure Bonds are safe, full investment back as bonds such as ICICI and IDBI
bonds
Infrastructure bonds can be pledged for loan
ICICI and IDBI Infrastructure Bonds can be purchased at Rs 5,000 each
5.3.4 Private Sector Debenture
Debentures are instruments meant for raising Long Term Debt Obligation. They have
following features:
Borrower pays Interest and Principal at specified times
Bank appointed as trustee and ensures firm fulfils contractual obligation
Secured by charge on immovable properties
Company chooses the coupon rate and redemption period
Carry call option to redeem debentures at certain price and put option to holder to seek
redemption at specific time at Predetermined Value
May have Convertible to Equity clause
5.3.5 Public Sector Undertaking Bonds
There are two broad varieties of PSU bonds as follows:
Taxable Bond – (i) Exempt only to certain limit according to section 80l
(ii) Interest Rate decided by PSU
Tax-free Bond – (i) Fully tax-exempt(ii) Interest Rate fixed by ministry of finance
Other important feature of PSU bonds are:
No Tax Deduction at Sourse
Transferable by mere endorsement and delivery
No Stamp Duty at transfer
Traded on Stock Exchange
5.3.6 Preference Shares
Preference Shares represent a hybrid security that partakes some characteristics of equity shares
and some attributes of debentures. The salient features of preference shares are as follows:
Characteristics of Equity share and debenture
Carry a fixed rate of Dividend
Dividend payable only on Distribution of Profit
Dividend is Cumulative
Preference shares are redeemable period of 7 to 12 years
Preference Dividend is Tax-Exempt up to certain limit
5.4 Financial Derivatives: Futures
A futures contract is an agreement between two parties to exchange:an asset for cash at a
predetermined future date for a price that specified today .It has following features:
Agreement to exchange asset at predetermined future date for specified price
Long Position: Party agree to purchase
Short Position: Party agree to sell
Long Position benefits from price increase,Short Position loses if price increases visa
versa
5.5 Financial Derivatives: Options
An option gives its owner the right to buy or sell an underlying asset on or before a given
date at a predetermined price. It has following features.
Buy of sell equity before a given date at Predetermine Price
No obligation for option holder
Call Option: Give option holder right to buy before certain date and exercise price
Put Option: Give right to option holder to sell a fixed number of share at exercise price
before given date
5.6 Life Insurance
The basic customer needs met by life insurance policies are protection and saving. We have
discussed some common Life Insurance Policies in this section. We have mentioned some
important features of various life insurance policies.
Endowment Assurance Policies
Participating Endowment Assurance
Guaranteed amount at maturity date or on death
One time or regular premium payment
Surrender of policy may be allowed
Can keep policy with out further premium but reduced assured sum
Can be used as security for the loan up to 90%
Endowment Assurance Policies
Non-Participating Endowment Assurance
Payment of bonus enhance the initial sum assured
Bonuses usually paid at the maturity or death of person
Option of using bonus for premium
Rest structure similar to Participating Endowment
Money Back Plan
Saving cum protection Policy
Repayment of certain percent of assured sum
Guaranteed annual addition(GAA)
Can be participating as well as non participating
In Participating GAA is replaced by bonus
No loan facilities attached
Whole life Assurance
Payable at death of person when ever it occur
Primarily for long term financial benefit to dependents
Benefits are paid on withdrawal of policy
Can take loan of 90% of surrender value
Both participating and non-participating version of policy
Term Assurance
Protection policy,benefit on death within specified term
Premium paid over the term or single outset
No payment or some proportion of premium as payment if policy holder survive
Death benefit at lower cost
Income to family of decrease till policy ends
Repay balance out standing under the loan
Immediate Annuity
Regular income after retirement
Policy could be structure for income for limited period
Single premium at inception
No payment on withdrawal
Can in non-participating as well as participating form
No surrender value of the option associated with the policy
Deferred Annuity
Regular Premium till vesting date
Premium buy income to be paid after vesting date
Helps to build Pension that become payable at retirement
Annuity can be paid on vesting date thus source to pay of other loan
Riders
Are add-ons to Insurance Policies,can be purchase with small premium
Common add on
Accidental Death Benefit(ADB) RiderCritical Illness (CI) RiderWaiver of Premium(WoP) Rider
Term Rider
5.7 Mutual Funds
Mutual funds are vehicle for collective Investment. There are three central entities to mutual
funds, The Sponsor who establishes the Mutual Fund and Asset Management Firm, The Mutual
Fund which Floats scheme where investors participates and The Asset Management Company
which manages the funds of Mutual Funds under various schemes.
5.7.1 Mutual Funds Open end /Close end scheme
Mutual fund operates in two modes Open end and close end. Salient Features open end and
close end are as follows.
Open End
Accepts fund on continuous basis
Withdrawal on continuous basis under repurchase
Has no maturity period
Not listed on Secondary Market
Close End
Subscription open for limited period
Continuous withdrawal not allowed
Has maturity period ( 5- 15 years)
Listed in Secondary Market
Common differences:
Closed end sold on discount over their Net Asset Value(NAV).Open-end schemes
valued close to it NAV
At redemption entire closed end investment liquidated
(i) Full value of investment not realised
(ii) Tax liability tend to be greater
No liquidity for Close end scheme thus hence efficient management and better
performance
Closed end sold on discount over their Net Asset Value(NAV).Open-end schemes
valued close to it NAV
At redemption entire closed end investment liquidated
(i) Full value of investment not realized
(ii) Tax liability tend to be greater
No liquidity for Close end scheme thus hence efficient management and better
performance
5.7.2 Mutual Funds: Equity Schemes and Debt Scheme
Instead of directly buying the equity give it to Mutual Funds who invest for you in Equity
schemes, Debt schemes, and balanced schemes. Various sections of Equity scheme and Debt
Schemes are as follows:
Equity Schemes
Growth Index: 80 - 95% invested in equity and related instrument. Objective to achieve long
term capital growth
Index Schemes: Invested in equity stock that comprises given stock Market. Thus profit moves
with the movement of index
Sectoral Schemes: Invests in specific sectors like pharmaceuticals, IT. Appeals to investor
interest
Debt Schemes
Income Scheme: Invested in fixed income securities as Government of India securities, state and
local government, Money market instrument .Small portion in equity instrument. Aim for steady
income
Government Security Scheme: Invest in sovereign securities of central and state government.
Aim modest return without any credit risk
Money Market Scheme: Invested primarily in money market scheme like treasury bills ,
commercial paper, certificates of debt etc. Aim for liquidity, periodic income
Balance Schemes
Distribute corpus across equity and debt in balance manner. This help in distributing the risk
hence suited mostly by service sector people earning regular income.
Chapter 6
Intelligent System for Investment Portfolio
As discussed in previous chapter Expert System, CBR and Genetic Algorithm have some
advantages and some disadvantages. These algorithms can be prudently combined to from
Intelligent System which utilizes unique features of all the three techniques and maximizes the
effectiveness of the system.
In Intelligent System first Expert system is use to take then input from the user and produce
some relevant input case structures. These are input to CBR (figure 6.1) which gives suitable
investment option which are also stored in form of cases. Then we use Genetic Algorithm to
determine the division of allotted fund in the given portfolio which will results in maximum
return or balance the risk profile based according to input specification.
Figure 6.1: Model for Applying ES, CBR and Genetic Algorithm for investment Portfolio
6.1 Combining Expert Systems and CBR
The architecture is intended for domains that are understood reasonably well, but still
imperfectly. It uses a set of rules, which are taken to be only approximately correct, to obtain a
preliminary answer for a given problem; it then draws analogies from cases to handle exceptions
to the rules. Having rules together with cases not only increases the architecture's domain
coverage, it also allows innovative ways of doing case-based reasoning: the same rules that are
used for rule-based reasoning are also used by the case-based component to do case indexing and
case adaptation.
Having rules together with the cases not only allows the architecture to take advantage of
more domain knowledge, it also allows innovations in CBR technology. The architecture
incorporates novel methods for CBR that are based on exploiting the rules of the RBR
component [27]. The rules are used to index the cases. The indexing scheme, termed predictionbased indexing, hangs cases directly of the rules, using the rule antecedents to supply appropriate
cues for case retrieval. This avoids having to analyze the domain to identify a suitable
vocabulary of direct and derived indexing features; instead, it takes advantage of the domain
structure implicit in the rules, and hence already available.
The core method is the heart of the architecture; it is the part that solves problems by a
combination of RBR or ES and CBR. The central idea is to apply the rules to the target problem
to get an approximate answer, and to draw analogies from cases to cover exceptions to the rules.
The indexing scheme, termed prediction-based indexing, hangs cases directly o the rules,
using the rule antecedents to supply appropriate cues for case retrieval. This avoids having to
analyze the domain to identify a suitable vocabulary of direct and derived indexing features;
instead, it takes advantage of the domain structure implicit in the rules, and hence already
available.
6.2 Combining CBR and Genetic Algorithm
After running CBR our output will be several investment options. Now we utilize important
property of Genetic algorithm that is optimization. Genetic Algorithm will be used to ration of
the money in various investment options which will optimize requirement stated by initial inputs
which may be risk, return, liquidity, tax exemption etc.
Chapter 7
Conclusion
The studies in the first stage were more of the explorative nature in terms of having a look at
all the available literature related to the subject in question. The overlaps and distinctness of the
material studied has helped in building a comprehensive picture of the various concepts
involved.
Various Algorithms of Artificial Intelligence are playing important roles in various decisions
making in finance. Expert System presents unique method to use rules that govern various
process outcomes. Case Base Reasoning helps in utilizing previous knowledge and experience in
the field in solving current problem. Genetic Algorithm helps in optimization of recourses. Over
all using all three techniques can develop an Intelligent System that will help investor to manage
there investment portfolio with maximum benefit.
7.1 Possible route for the Second stage
As a continuation of study from the first stage, the second stage would try to get in-depth
understanding of Investment Alternatives and discuss about various variable according to which
these investment could be identified. This information will be used to build the case structure and
indexing of cases to be used for CBR. It would then derive the rules for expert system by which
queries could be formed from input user information. Implementation and testing of Intelligent
System which combine ES,CBR and GA will be the next step of the project.
Reference:
[1] Brule, James F. Artificial Intelligence: Theory, Logic and Application, 1986, Blue Ridge
Summit, PA: TAB Books.
[2] Harmon, Paul and King, David. Expert Systems: Artificial Intelligence in Business. 1985,
New York: Wiley.
[3] Winston, Patrick H. and Prendergast, Karen A. (Editors). The AI Business: Commercial
Use of Artificial Intelligence, 1984, Cambridge, MA: The MIT Press.
[4] R. Bergmann et al. (Eds.): INRECA-Methodology 2nd Edition, LNAI 1612, pp. 9–20,
2003,Springer-Verlag Berlin Heidelberg 2003
[5] Liebowitz, Jay, Introduction to Expert Systems, 1988, Santa Cruz, CA: Mitchell Publishing,
Inc.
[6] Goldberg, David E (1989), Genetic Algorithms in Search, Optimization and Machine
Learning, Kluwer Academic Publishers, Boston, MA.
[7] Goldberg, David E (2002), The Design of Innovation: Lessons from and for competent
Genetic Algorithms, Addison-Wesley, Reading, MA. (Manago et al. 1994, Bergmann
2002).
[8] Jay Liebowitz, Worldwide Perspectives and Trends in Expert Systems, 1997, American
Association for Artificial Intelligence
[9] Engelmore S Robert, Feigenbaum Edward Introduction to Knowledge Base System
Published: May 1993; WTEC Hyper-Librarian
[10] Brown E Carol, O'Leary E Daniel , Introduction to Artificial Intelligence and expert
system, Copyright 1993, 1994, 1995
[11] H. Penny Nii, Tools and Infrastructure for Knowledge-Based Systems, May 1993; WTEC
Hyper-Librarian
[12] Luger George Artificial Intelligence: Structures and Strategies for Complex Problem
Solving, fifth edition 2004
[13] Negnevitsky M. Artificial Intelligence. A Guide to Intelligent Systems, Addison Wesley,
Second Edition, 2005
[14] Introduction to Expert Systems Copyright © 2001 by eXpertise2Go.com.
[15] Berghofer R Thomas, Knowledge Maintenance of Case–Based Reasoning Systems, 2003
Akademische Verlagsgesellschaft Aka GmbH, Berlin.
[16] Aamodt A. and Plaza E. Case based reasoning: Foundational issues, methodological
variations, and system approaches. AI Communications, 7(1):39.59, 1994.
[17] Mántaras1 R, Bridge D, McSherry D Retrieval in Case-Based Reasoning: An Overview
[18] D.W. Aha, L.A. Breslow, and H. Muñoz-Avila (2001), Conversational Case-Based
Reasoning. Applide Intelligence 14 (1), 9-32.
[19] Retrieval. Proceedings of the 2nd International Conference on Case-Based Reasoning
(Eds. D. B. Leake and E. Plaza), Springer-Verlag, 279-290.
[20] P. Cunningham, D. Doyle and J. Loughrey (2003), An Evaluation of the Usefulness of
Case-Based Explanation Proceedings of the 5th International Conference on Case-Based
Reasoning (Eds. K.D.Ashley and D.G. Bridge). Springer, 122-130.
[21] K.D. Forbus, D. Gentner, and K. Law (1994), MAC/FAC: A Model of Similarity-based
Retrieval, Cognitive Science 19 (2), 141-205
[22] D. Wettschereck and D. Aha (1995), Weighting Features. Proceedings of the 1st
International Conference on Case-Based Reasoning (Eds. M. Veloso and A. Aamodt),
Springer-Verlag, 347-358
[23] Agnar Aamodt , Enric Plaza, Case-based reasoning: foundational issues, methodological
variations, and system approaches, AI Communications, v.7 n.1, p.39-59, March 199
[24] Mitchell,
Melanie, (1996), An Introduction to Genetic Algorithms, MIT Press, Cambridge,
MA.
[25] Vose, Michael D (1999), The Simple Genetic Algorithm: Foundations and Theory, MIT
Press, Cambridge, MA.
[26] Prasanna Chandra Investment Analysis and portfolio Management, TMH
[27] Golding R. Andrew, Rosenbloom Paul (1995), Improving accuracy by combining rulebased and case-based reasoning, TR-94-19a